ORNL data scientists are creating the ability to understand, interpret, query and model data generated by large and distributed information systems through the development of advanced architectures, toolsets, and interfaces. Over the years, as storage has become less expensive and computational capacity has increased, the challenge of real-time inference and data-driven strategic intelligence extraction is fast translating as the ability of data scientists to quickly digest diverse sources of seemingly-related information from multiple data assets. The bottleneck is not anymore about how to store and how much to store but it is about how datasets in different formats (structured, unstructured), hosted across different infrastructures (mainframes, cloud, custom-hardware, etc.), in different databases (row-oriented, column oriented, file-oriented etc.), and in different schemas can all be quickly and seamlessly presented, explored and analyzed. We provide research into emerging data architectures and processing methods for scientific, defense, and security applications with special requirements for linking real-time systems and heterogeneous databases.